Meter / Runtime Reading Tracker: Catch Bad Readings and Surface Assets Due for Service
Capture run-hours, cycles, or mileage per asset on every visit, validate each reading against the last one to catch impossible jumps, rollovers, and resets, and flag the assets creeping up on a usage-based service threshold — with a planner approving what to act on before anything is triggered.
A web tool where a tech logs a meter reading for an asset on each visit, the tool validates it against the prior reading (catching rollbacks, impossible jumps, meter rollovers and resets), computes usage since the last service, flags assets approaching their service threshold, and a planner reviews the flagged readings and due-soon list and approves which to act on. Readings and the due list export to CSV.
Before you start
- A Supabase account (free)
- A Vercel account (free)
- A Resend account (free)
- A CSV or Google Sheet of your assets with their prior readings and service thresholds
- Claude Code or any AI coding agent
The problem this kills
Your techs jot meter readings on a clipboard or punch them into a spreadsheet — run hours on a generator, cycles on a press, mileage on a fleet vehicle. Then the numbers go sideways. Someone fingers an extra zero and a forklift "gains" 90,000 hours overnight. A reading comes in lower than last time and nobody notices the meter rolled over at 9,999 or got swapped for a new one. Worst of all, the whole point of tracking usage — knowing when each machine is due for service — gets lost, because nobody is sitting there doing the subtraction across hundreds of assets with different units and different intervals.
So service gets driven by the calendar or by whoever shouts loudest, machines run past their hours, and warranty claims get denied because the usage record is a mess. The fix isn't fancy: capture each reading where the tech already is, check it against the last good one the moment it's entered, and quietly keep a running tally of usage so the assets creeping up on their threshold raise their hand. You do not need to be a developer to build a tool that does exactly that.
What you'll build
A simple internal web tool for your techs and maintenance planners. A tech logs a meter reading for an asset on each visit — run hours, cycles, mileage, whatever that asset uses — and the tool validates it against the prior reading on the spot: it catches readings that go backwards (a rollback), jumps that are physically impossible for the time elapsed, and it knows the difference between a real meter rollover (the counter wrapped past its maximum) and a meter reset (a new or replaced meter starting from zero). It computes usage since the last service for every asset and flags the ones approaching their service threshold ("due soon" / "overdue"). A planner reviews the flagged readings and the due-soon list and approves which to act on before any service is triggered. Readings and the due list export to CSV any time.
What's inside the Implementation Plan
The downloadable plan is a step-by-step file you paste into an AI coding agent. It opens by interviewing you about your business — what you meter and in which units per asset, how readings reach you today, the real column names in your asset and reading sheets, your service thresholds and intervals per asset class, what a plausible jump looks like for each meter type, how you tell a rollover from a reset, your typical and peak reading volumes, who is allowed to approve acting on a flagged reading, and the messy edge cases (swapped meters, estimated readings, multiple meters on one machine). It reads a short spec back to you for a thumbs-up, then builds the tool around your assets, your units, and your rules instead of a generic template. From there it walks the agent through the data model, the reading-capture flow, the validation engine, the due-soon calculation, the planner-approval gate, and the CSV exports. Every step ends with a ready-to-copy prompt.
The governance it includes (this is the point)
This isn't a toy. The plan builds in the controls a real asset function needs: login so only your team can use it, row-level security so people only see their own organization's assets and readings, a complete audit trail of every reading, validation result, and decision (who, what, when), a hard human-approval gate so no flagged reading or due-soon asset triggers service until a planner signs off, and duplicate guards — a reading is deduped on (asset + reading date) so the same visit can't be logged twice and accidentally double-count usage. The tool exists to make a careful human decision easy: the AI flags the suspicious reading and the due-soon asset, a person decides what happens next.
Who it's for
Field-service technicians who capture readings in the field, maintenance planners who decide what gets serviced, and asset controllers who own the usage record and the warranty paperwork. If you can describe what your machines meter and when they're due, you can build this.
You've got this — open the plan, paste the first prompt, and you'll be catching your first bad reading by the end of the afternoon.